The role of blood platelets in nucleoside metabolism: assay, cellular location and significance of thymidine phosphorylase in human blood

Mutation Research, 200 (1988) 99-116

99

Elsevier MTR02308

The role of blood platelets in nucleoside metabolism: assay, cellular location and significance of thymidine phosphorylase in human blood T. Shaw, R.H. Smillie a and D.G. MacPhee Department of Microbiology, LaTrobe University, Bundoora 3083 (Australia) and a Department of Biochemistry, University of Melbourne, Parkville 3052 (,4ustralia) (Received 27 January 1988) (Accepted 24 February 1988)

Keywords: Human blood; Platelets; Thymidine phosphorylase; Nudeoside metabolism; HPLC

Summary The enzyme thymidine phosphorylase (thymidine : orthophosphate deoxyribosyltransferase, EC 2.4.2.4), which plays a crucial role in nucleic acid metabolism in both prokaryotic and eukaryotic cells by regulating the availability of thymidine, is present in mammalian blood. Here we describe a simple, rapid HPLC-based micromethod for the assay of blood thymidine phosphorylase. We have arbitarily defined 1 unit of blood thymidine phosphorylase activity as the activity required to produce a 1-nM increment in the plasma concentration of thymine after incubation for I h at 37 °C with a saturating concentration of exogenous thymidine. In normal adults, whole (peripheral venous) blood thymidine phosphorylase activity with blood cells intact was 64 _+ 11 units (mean _+ S.D., n = 20, range 45-89). The apparent Michaelis constant for thymidine was of the order of 10 - 4 M but varied nearly 5-fold between different individuals. Activity increased when blood cells were permeabilised or lysed with non-ionic detergents, implying that thymidine phosphorylase is an intracellular enzyme which may be influenced by exogenous as well as intracellular factors. When blood from normal donors was fractionated, thymidine phosphorylase activity consistently co-isolated with platelets. Whole-blood thymidine phosphorylase activity correlated well with platelet parameters. Although thymidine phosphorylase activity was also detected in plasma and serum, the small size and notorious fragility of platelets suggest its platelet origin. Blood from leukaemic donors showed significantly increased thyrnidine phosphorylase activity compared to normal controls (mean activity _ S.D. was 96 _+27 units; range 58-140, n = 8). Thymine formation from thymidine was temperature- and pH-dependent in whole blood. 2'-Deoxyuridine and 3 of its 5-halogenated analogues (but not 3'-azido-3'-deoxythymidine (AZT), were catabolised by blood thymidine phosphorylase, even during blood clotting at room temperature. Assumptions about in vivo concentrations of these compounds should therefore be interpreted cautiously. In the presence of high concentrations of thymine and suitable deoxyribose donors, small amounts of thymidine were formed in some blood samples, so it is conceivable that thymidine catabolism may be reversible in vivo under some circumstances. Correspondence: T. Shaw, Department of Microbiology, LaTrobe University, Bundoora 3083 (Australia).

Abbreviations: K 2 EDTA, ethylene diamine tetra amino acetic acid dipotassium salt; HIV, human immune deficiency virus. Other abbreviations as defined in text. For convenience, the term 'cells' is used to describe all cell-derived blood components including platelets and erythrocytes, which are uncharacteristic of normal eukaryotic cells. 0027-5107/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)

100 Taken together, these data suggest that (i) blood platelets play an important role in thymidine homeostasis and may thereby significantly influence whole-body DNA metabolism; (ii) blood platelet functions are probably a major determinant of the efficacy of many important antineoplastic and antiviral nucleoside analogues and (iii) a platelet role in pyrimidine deoxynucleotide homeostasis may help to explain the well-documented but poorly-appreciated association between platelets and cancer.

Nucleoside salvage may be regarded as an energy economy of value to both prokaryotes and eukaryotes (Kornberg, 1980). The salvage pathway can be simplified to 2 steps: (i) carrier-mediated transport across the cell membrane and (ii) intracellular entrapment resulting from phosphorylations catalysed by specific kinases (Plagemann and Wohlhueter, 1980). In mammalian cells, the reaction catalysed by thymidine kinase is regarded as being particularly important because (i) it is cell-cycle-dependent and essentially irreversible, (ii) thymidine triphosphate (dTTP), the end-product of the thymidine salvage pathway, regulates not only thymidine salvage (by inhibition of thymidine kinase) but also production of other deoxyribonucleosides (by allosteric effects on ribonucleotide diphosphate reductase (Reichard, 1978)) and (iii) dTI'P may also influence DNA polymerase activity by interacting with a regulatory protein (Steinberg et al., 1979). Moreover, the possible genetic consequences of both thymidine starvation and thymidine excess encompass the entire spectrum of possible genetic lesions, and have been well documented (for recent reviews see Barclay et al., 1982; Meuth, 1984; Haynes, 1985; Haynes and Kunz, 1986). Fig. 1 illustrates the alternative metabolic fates for thymidine in mammalian cells. When extracellular thymidine concentrations are in the micromolar range, thymidine kinase activity (which is energy-dependent) rather than uptake (which is energy-independent) is probably rate-limiting for DNA synthesis (Barlow and Ord, 1975). Paradoxically, thymidine does not accumulate intracellularly and cellular thymine nucleotide pools are extremely small during most of the cell cycle (Barlow and Or(], 1975; Usher and Reiter, 1977; Bjursell and Skoog, 1980); this implies that catabolism may regulate the availability of thymine nucleotides by competing with thymidine kinase for substrate (Cooper et al., 1972;

Usher and Reiter, 1977; Bjursell and Skoog, 1980). In support of this postulate there is convincing evidence to suggest that in mammalian tissues

thymine

extrac@llularspace

thymidine

I thymine~

cytosoJ

,\

TOP

IMP

1

saluage catabolism de n o u o synthesis

Fig. 1. Thymidinemetabolismin mammaliancells (simplified). Steps or enzymes: 1, uptake; 2, diffusion or excretion; 3, thymidine phosphorylase; 4, thymidylatesynthase; 5, thymidine kinase; 6, thymidylatekinase; 7, nucleosidediphosphate kinase; 8, DNA polymerase(s);9, nucleases, 10, nucleotidases. Note that thymine is normally lost by diffusion but may be reduced to dihydrothymineand further catabolised in some cells. Reduction to dihydrothymineis rate-limiting for thymidine catabolism and is quantitatively unimportant in most tissues. The assay for blood thymidinephosphorylase activity depends on measuring thymine released into the extracellular space.

101

which salvage thymidine the size of the intracellular thymine nucleotide pool is directly proportional to the exogenous thymidine concentration (Cooper et al., 1966). Further support is provided by a recent observation than in at least 1 cell system in vitro, the number of S-phase nuclei (as judged by incorporation of [3H]thymidine into DNA) increases in parallel with the exogenous thymidine concentration (Puzas and Brand, 1986). Two pyrimidine nucleoside phosphorylases, uridine phosphorylase (EC 2.4.2.3) and thymidine phosphorylase (EC 2.4.2.4), are known to initiate thymidine catabolism in mammals (Friedkin and Roberts, 1954; Friedkin and Kalckar, 1961; Kraut and Yamada, 1971). They catalyse the reversible phosphorylysis of their substrates as follows: pyrimidine(deoxy)ribonucleoside + Pi pyrimidine base + (deoxy)ribose-l-phosphate The 2 enzymes differ in their pH optima, tissue distribution and substrate specificity (Friedkin and Roberts, 1954; Friedkin and Kalckar, 1961; Kraut and Yamada, 1971); only thymidine phosphorylase can also catalyse direct pentosyl transfer from deoxyribonucleosides to pyrimidine bases (Zimmerman, 1964; Gallo and Perry, 1969): pyrimidine base 1 + deoxyribonucleoside2 --* pyrimidine deoxyribonucleoside1 + base 2 Blood thymidine phosphorylase activity, originally observed in leukocytes by Marsh and Perry (1964), was subsequently detected in platelets by Schneider et al. (1973) and in plasma by Pauly et al. (1977, 1978), although Woodman (1979) has presented evidence that plasma thymidine phosphorylase activity is the result of cell lysis. More recently, it has been recognized that platelets can account for a substantial proportion of total blood thymidine phosphorylase activity (Desgranges et al., 1981; Pero et al., 1984; Shaw, 1986; Shaw and MacPhee, 1986; Bodycote and Wolff, 1986). We consider detection of thymidine phosphorylase activity in blood to be important because (i) thymidine phosphorylase appears to have

the potential to degrade or produce thymidine depending on its environment and may therefore have the capacity to restore thymidine imbalances; (ii) blood appears to be the major source of thymidine and other pyrimidine nucleosides for many mammalian tissues (Weismann, 1966; Levine et al., 1974 and references cited therein); (iii) mammalian blood thymidine has a very short half-life (8-10 min) (Rubini et al., 1960; Ensminger and Frei, 1977) and normal blood concentrations appear to be stringently maintained in the submicromolar range (Hughes et al., 1973; Nottebrock and Then, 1977; Dudman et al., 1981; Holden et al., 1980) despite the high levels which could theoretically be achieved as a result of cell turnover (see Appendix 1); (iv) abnormally high serum thymidine concentrations, presumably due to increased cell turnover, have been observed following X-irradiation (Hughes et al., 1973) and also in individuals suffering from certain types of leukaemia and haemolytic anaemia (Holden et al., 1980); but they may equally well be due to impaired catabolism; (v) it is difficult to maintain high plasma concentrations of thymidine even by intravenous infusion of massive doses (Ensminger and Frei, 1977; Kufe et al., 1980; Martin et al., 1980; Berlinger et al., 1987; O'Dwyer et al., 1987), and finally (vi) blood thymidine phosphorylase almost certainly contributes to the rapid clearance of thymidine and many deoxyuridine analogues which are in use as antineoplastic and antiviral agents; it may also be involved in their regeneration (Desgranges et al., 1986). Although catabolism of these compounds has frequently been assumed to be an almost exclusively hepatic function (see, e.g., Myers, 1981; Berlinger et al., 1987), it has also been observed that their rates of plasma clearance exceed cardiac output, indicating that other tissues besides the liver must be involved (Martin et al., 1980; Myers, 1981; O'Dwyer et al., 1987). To date, most methods for assessing thymidine phosphorylase activity in complex biological samples have been based on detection of labelled catabolites after extraction and separation by thin-layer chromatography (TLC) (e.g., Schneider et al., 1973; Usher and Reiter, 1977; Pauly et al., 1977, 1978; Woodman, 1979; Pero et al., 1984). These methods are both slow and tedious, and

102 their results are difficult to quantify, let alone standardise. Recent advances in high pressure liquid chromatography (HPLC) have resulted in the development of excellent methods for separation and detection of nucleosides, nucleobases and their derivatives in biological samples (Hartwick et al., 1979a,b; Taylor et al., 1980; review: Simpson and Brown, 1986); it seemed likely that the development of an HPLC-based assay for thymidine phosphorylase activity in minimally treated samples might therefore be feasible. Moreover, the known presence of thymidine phosphorylase activity and extremely low thymidine and thymine concentrations in mammalian blood suggested that the assay of blood thymidine phosphorylase might be simpler and more meaningful than measurements of basal levels of its substrate or product. Here we describe a simple assay for blood thymidine phosphorylase, and report upon normal levels and cellular location of enzyme activity in peripheral human blood. We also speculate on the probable in vivo importance of thymidine phosphorylase in relation to nucleotide metabolism and deoxynucleotide pool balance. Materials and methods

were prepared by mixing anticoagulated blood with an equal volume of 2% (w : v) aqueous Triton X-100. Additional blood samples were collected into tubes without anticoagulant (but in some cases with added nucleosides), and allowed to clot for 1 h at room temperature, after which serum was removed.

Pathological blood samples Small samples of freshly drawn K2EDTA-anticoagulated blood were obtained from leukaemic individuals who had attended an out-patient clinic for routine haematological tests. These were kindly made available by Dr M.G. Whiteside of the Cabrini Medical Centre, Malvern, Victoria, who also provided the relevant cell counts, but not medical histories.

Substrate preparation Aliquots (100 /~1) of ethanolic thymidine solutions at concentrations which ranged from 100 to 1000 # M were dispensed into sterile 1.5-ml microfuge tubes. The tubes were placed in a heating block at 45 ° C until the ethanol had completely evaporated, then sealed and stored at room temperature until needed. Other nucleosides were similarly treated.

Materials Chemicals and solvents were analytical or HPLC grade. Nucleosides and nucleobases were from Sigma or Boehringer; Percoll was supplied by Pharmacia. Methyl[3H]thymidine (code TRK.686, specific activity 70 Ci mmole -1) was purchased from Amersham International.

Blood collection and sample preparation Blood donors were apparently healthy drug-free adult volunteers whose ages ranged from approximately 20 to 60 years and whose blood cell counts and serum biochemistry were within the normal range. The same donors were used on several separate occasions over a 6-month period. One (female) donor who had undergone splenectomy about 12 months before blood samples were first taken had a slightly elevated platelet count but her haematological and biochemical values were otherwise normal. Blood was obtained by antecubital venepuncture and immediately anticoagulated with 1.5 mg m1-1 K2EDTA. Lysates

Assay method Substrate-containing tubes were prewarmed to 37 ° C in a heating block before addition of 100 #1 of anticoagulated blood or other sample for assay. Incubation at 37 ° C was continued for up to 5 h. At the end of the incubation period tubes were spun in a microfuge for 60 s at 14 000 x g to pellet cells and cell debris. Aliquots (25 #1) of the supernatant were removed and vortex-mixed in fresh microfuge tubes with 100/~1 methanol. Each tube was centrifuged for 10 min at 14000 x g immediately before removal of the deproteinised supernatant for analysis by HPLC. Controls lacking substrate were treated identically. Fig. 2 summarises the assay method.

Factors affecting thymidine phosphorylase activity and cellular location of thymidine phosphorylase activity Conditions were modified to test the effects on the assay of temperature, pH, non-ionic deter-

103

f

evaporate solvent

1O0I.tl methanolic thymidine solution

RAPIDMICROASSAY forBLOOD THYMIDINPHOSPHORYLASE E

methanol

1O0~1bloodsample 25I.tl plasma ~

j' 37oc

~

14,000g ~min~- 1~

1hour

J

mix

/4,ooo

v Fig. 2. Summary of method for rapid assay of thymidine phosphorylase activity in human blood.

gents, blockage of nucleoside uptake and sample storage time. Some experiments required preincubation without substrate. In these experiments thymidine was added in solution (10 #1 of 5 mM solution in isotonic saline per 90-#1 sample). Details are given with results.

Chromatography: HPLC Deproteinised samples were injected into a Varian model 5000 liquid chromatograph using a Rheodyne injector fitted with a 10-#1 sampling loop. A reversed-phase ClS analytical column (Varian MCH-5 N-CAP, particle size 5 #M, dimensions 150 × 4 mm), protected by a guard column containing the same solid phase, was used to separate sample components which were eluted isocratically at ambient temperature and a flow rate of 1.0 ml min-1. The mobile phase was 100 mM formic acid adjusted to pH 4.0 with ammonia solution and contained 5 mM NaC1 and 5% (v : v) aqueous methanol. Thymidine and thymine were detected by UV absorption at 266 nm using a Varian UV 100 detector. A coupled integrator (Varian model 4720) permitted automatic calibra-

tion and peak analysis. The analytical column was regularly regenerated by overnight flushing with methanol. Guard columns were replaced after about 200 analyses. It is worth mentioning that we were able to use a single (Varian MCH-5 N-CAP) analytical HPLC column for nearly 7000 thymidine phosphorylase assays over a 12-month period without significant loss of resolution. Thin-layer chromatography (TLC) was performed as previously described (Shaw and MacPhee, 1986).

Blood fractionation Percoll density gradients were used to separate blood cells. The procedure used was essentially a miniaturisation of the technique described previously (Shaw and MacPhee, 1986). Percoll was prepared as described by Vincent and Nadeau (1984) so as to produce a sterile isotonic medium of density 1.077-g cm -3, containing 140 mM NaCl, 5 mM KC1, 1 mM EDTA, 10 mM Tris-HC1, pH 7.4 (final concentrations). Aliquots (1.1 ml) of this medium were dispensed into sterile 2.5-ml microfuge tubes and then carefully overlaid with

104

1.0 ml anticoagulated blood. A brief centrifugation (60 s, at approximately 10000 × g without brake) in a microfuge fitted with a swinging bucket rotor banded platelets and some small lymphocytes at the gradient interface. After centrifugation, gradients were divided into 3 fractions of equal volume (see Fig. 8). Platelet recoveries and contamination of the platelet/small lymphocyte fraction with other cells were comparable with those previously achieved (> 95% platelet recovery relative to whole blood and negligible red cell contamination; see Shaw and MacPhee, 1986). Using this method, blood samples could be fractionated within 5 min of collection. Each fraction was assayed for thymidine phosphorylase activity.

I"

20 minutes

-I a

b thymidine

d

pthymine Results

Figs. 1-12 illustrate results. Each experiment was repeated several times, but in most cases results from single representative experiments are shown because of the rather wide inter-subject variation in blood thymidine phosphorylase activity and characteristics.

Chromatography: HPLC Conditions were optimised for the rapid separation of thymine and thymidine from other sample components which absorbed at 266 nm, a wavelength chosen as a compromise between absorbance maxima of thymine and thymidine under acidic conditions (Dawson et al., 1986). Compounds separated by HPLC were tentatively identified by the standard addition technique using known standards together with reference to published data (Hartwick et al., 1979a,b; Taylor et al., 1980) (see Fig. 3). Detector response was linear and reproducible (invariably within + 4% for sample concentrations greater than 10 #M) throughout the entire experimental range. After a single calibration, the response to the same standard preparation (containing 100 #M each of thymine and thymidine) also remained within + 4% over a 12-month period. The absolute detection limit for both thymine and thymidine was about 0.5 pmole, equivalent to a concentration of 2.5 #M in the sample being analysed. Although the assay's sensitivity could be improved by increasing the sample volume and using serum or plasma ultrafiltrates

Fig. 3. Detection of thymidine and thymine in methanolic extracts of plasma by HPLC. (a) HPLC analysis of methanolic extract of normal plasma (no incubation or added substrate). Peak 1: EDTA, adenine derivatives; other purine metabolites and uracil elute before peak 2, which has tentatively been identified as uridine, Compare with (d) and notice the absence of detectable thymine or thymidine. Dihydrothymine, if present, elutes ahead of thymine and thymidine in the region indicated by the arrow. (b) Same sample immediately after substrate addition. (c) Methanolic extract of plasma from a leukaemic individual. The thymine concentration in the extract was 13.1 #M corresponding to a basal plasma concentration of 65.5 #M. Compare with chromatograph from normal sample in (a) above. (d) Same sample as in (a) and (b) showing separation of thymine from remaining thymidine after incubation. When [methyl-3H]thymidine tracer was present during assay, label eluted only with the thymine and thymidine peaks.

rather than methanolic extracts, this was deemed unnecessary and too costly for routine work. Ultimately, the relative insensitivity of the assay to background proved to be advantageous (see Discussion). Extraction of normal plasma with methanol yielded few separable compounds which absorbed at 266 nm (Fig. 3a). Basal levels of thymine and thymidine in normal plasma were invariably below the detection limit of the assay. Normal plasma concentrations of free nucleoside and base must therefore have been less than 2.5 #M. Basal thymidine was detectable in 6 of 8 leukaemic blood samples. Plasma concentrations (#M) were: 2.5, 5.0, 7.0, 10.0, 12.5, 15.5. A basal

105

thymine concentration of 65 laM was detected in 1 leukaemic sample (see Fig. 3c). (The possibility that this peak was due to a drug such as 5-fluorouracil, which elutes in a similar position, was eliminated after consultation with the individual's physician.) HPLC analysis also revealed abnormal early-eluting peaks, presumably representing purine metabolites, in all leukaemic samples (also illustrated by Fig. 3c). Preliminary experiments (results not shown here) using [3H]thymidine tracer demonstrated that recovery of thymidine added to blood immediately before extraction was > 95%, and that transmembranous equilibrium was achieved in less than 20 s. The plasma concentrations of thymine and thymidine are therefore good indicators of intracellular levels, as Desgranges and colleagues (1981, 1983, 1986) have also found. Although we and others have previously detected dihydropyrimidine dehydrogenase activity in some preparations of human platelets (Pero et al., 1984; Shaw, 1986; Shaw and MacPhee, 1986), thymine catabolism in normal samples was rarely quantitatively significant during the first hour of incubation, because when [3H]thymidine tracer was included, neither TLC nor HPLC analyses showed significant peaks of radioactivity corresponding to dihydrothymine, which elutes ahead of thymine and thymidine in both chromatographic systems (see Fig. 3a). Unfortunately it is impossible to detect thymine catabolites by UV absorbance either because they do not absorb or because of unacceptable background due to the presence in plasma of numerous interfering compounds. (Derivitisation, another possibility, is impractical for the latter reason.) Addition of thymine to whole blood did not result in detectable thymidine production, but small amounts of thymidine could sometimes be produced from thymine if high concentrations of deoxyguanosine (or other suitable deoxyribose donor) were also present. By itself, 2-deoxyribose had no effect on the fate of thymine, but appeared to block thymidine uptake, since it inhibited catabolism (results not shown). Enzyme kinetics A remarkably high thymidine concentration (almost 3 orders of magnitude greater than reported plasma concentrations) was required to

5O 40 30

E

20

t-

10 I

0

200

!

!

400 600 [thymidine](I.tM)

800

-0.01 0.00 0.01 1/[lhyrnidine](laM )

0.02

0

0.06

0.05 0.04 r-

0.03

E c-

0.02

T--

0.01 0.00 -0.02

!

125 IO0

~

75

x:

50 25 0 0

1

2

3 hours

4

5

6

Fig. 4. Kinetics of TP activity in blood. (a) The standard assay performed using a range of substrate concentrations to produce a saturation curve. (b) Lineweaver-Burke plots of data presented in (a). (c) Thymine production in whole blood samples as a function of time.

saturate blood thymidine phosphorylase activity (Fig. 4a). Buffering of free thymidine by nonspecific binding to plasma proteins may have been partly responsible for this surprising finding.

106 T h y m i n e f o r m a t i o n in K 2 E D T A - a n t i c o a g u l a t e d whole blood showed M i c h a e l i s - M e n t e n kinetics at t h y m i d i n e c o n c e n t r a t i o n s below 500 # M (Fig. 4b). I n whole blood, the a p p a r e n t K m for t h y m i d i n e was of the order of 10 - 4 M, varying b e t w e e n different i n d i v i d u a l s (range 0 . 4 - 1 . 9 × 10 - 4 M, n = 5). The Vm~, for t h y m i n e f o r m a t i o n also varied b e t w e e n individuals (range 1 5 - 1 6 0 n M m1-1 h -1, n = 5). I n b l o o d lysates the c o r r e s p o n d i n g values were even more variable. If the d i l u t i o n factor was taken into account, it was clear that lysis always increases Vm~,, b u t did n o t alter K m in a consistent way. U n d e r the c o n d i t i o n s of assay, the rate of t h y m i n e f o r m a t i o n was a p p r o x i m a t e l y linear for the first h o u r b u t gradually decreased (Fig. 4c).

Catabolism of thymidine during clotting," activity in serum and plasma Because some t h y m i d i n e c a t a b o l i s m occurred at r o o m t e m p e r a t u r e (Fig. 5a), the possibility that catabolism could occur d u r i n g clotting was investigated. A s a t u r a t i n g c o n c e n t r a t i o n of thymidine was added to b l o o d samples which were

80 60 .

w

E

El

40

20

Factors affecting thymidine phosphorylase activity T h y m i n e f o r m a t i o n was m a x i m a l at p H 5.5-6.5 (Fig. 5a), a n d increased with temperature (Fig. 5b). T h y m i d i n e phosphorylase activity appeared to decay slowly if b l o o d samples were stored at r o o m t e m p e r a t u r e (Fig. 5c). However, provided that assays were set u p within a few hours of b l o o d collection, differences in the a m o u n t of t h y m i n e p r o d u c e d from a saturating dose of t h y m i d i n e were small.

__,w

0

20 30 temperature ~°'C)

40

50

18 16

/

14 rE

12 10

8 Fig. 5. Factors affecting blood TP assay. (a) Temperature. Samples were equilibrated for 15 rain at the temperatures indicated before substrate addition and incubation continued at the same temperature. (b) pH. Cell-free plasma was prepared from part of a blood sample. Very small quantities ( < 1% v : v) of I M lactic acid or 1 M sodium bicarbonate were added to separate plasma samples which were then mixed in different ratios. An equal volume of each of these preparations was mixed gently with an aliquot of the original blood sample 5 rain before addition of substrat¢. (Note that only half the usual number of cells were present in this assay.) Plasma pHs were checked at the end of the assay. The relatively small effect of extracellular pH changes may reflect a failure to influence intracellular pH significantly, so that observed changes may be due to effects on extracelhilar enzyme only. (c) Samplestorage. A freshly collected anticoagulated blood sample was split into a number of aliquots, which were kept at room temperature. Each aliquot was assayed at a different time after blood collection.

10

_

b !

6

!

6

4

8 pH

r,-

•- -

35

C ~0

T

0

2

!

T

!

!

4 6 8 10 hours after collection

T

12

107

Catabolism of pharmacologically important compounds

r = 0.66

O3

E

[]

30

Other pyrimidine nucleoside-base pairs could also be separated rapidly by HPLC using appropriate concentrations of methanol in the mobile phase. Catabolism of deoxyuridine (dUdR) and its 5-fluoro-, 5-bromo- and 5-iodo-derivatives

[] []m[]

oO9 i

y

E 10 10

m

[]

"

! 20 [thymine] - clotting at RT

El

30

Fig. 6. Thymine formation during d o t t i n g and thymidine phosphorylase activity in serum. Thymidine (500/tM) was added to blood which was allowed to clot at room temperature. Separate, paired cell-free plasma and serum samples were assayed for thymidine phosphorylase activity at 3 7 " C . Activity in serum at 37 o C was similar to activity during clotting at room temperature. Plasma values (not shown) were similar to those for serum.

5O

_¢:

25

plasmaonly !

0

allowed to clot at room temperature for 1 h, after which serum thymine concentrations were determined by HPLC. Serum and plasma collected from replicate blood samples were assayed for thymidine phosphorylase activity at 37°C. The substantial but variable amount of thymine which was produced during dotting at room temperature correlated more closely with serum activity than with blood cell counts (Fig. 6). Significant differences in activity between paired plasma and sera were not detected (plasma results not shown).

1.0

80

b

6O E

40 20 0 0

Fig. 7. Cellular location of thymidine phosphorylase activity. (a) Effect of cell number. Halving dilutions of a blood sample were prepared using autologous plasma as the diluent. Each dilution was assayed for activity. Thymine production was directly proportional to the number of cells present. The dotted line indicates the level of plasma activity. (b) Effect of membrane disruption. The non-ionic detergent Triton X-100 caused a concentration-dependent increase in thymine formation production when added to whole blood. Lysis was evident at high concentrations. In this experiment the detergent was diluted with isotonic saline. Similar results were obtained with Nonidet P40 and Tween 80. (c) effect ofdipyridamole. Halving dilutions of the nudeoside uptake blocker dipyridamole solution were prepared in isotonic saline. Whole blood or lysates were preincubated at room temperature with different dipyridamole concentrations for 15 min before addition of substrate.

I

O.5 bloed:plasma ratio

I.O

I

l

I

I

2

3

4

[Triton X-1001 (% w:v) 9O

lysates

7O

._ E

C

I

5O

~

30

.

~

cellsintact i

10

!

0

2

I

I

I

4 6 8 [dipyridamole](~.g/ml)

10

108 and of 3'-azido-3'-deoxythymidine (AZT) was investigated, both in anticoagulated blood at 37 ° C (normal assay) and during clotting at room temperature (about 22 ° C). After 1 h incubation with 500 /tM substrate, plasma and serum base concentrations varied over a wide range (equivalent to 18-140 assay units) for the deoxyuridines, depending on the 5-substituent and the individual blood sample. Generally, the extent of catabolism was I U d R > BrUdr > (dThd) > F U d R > UdR, suggesting that uptake and enzyme affinity may be directly related to the size of the 5-substituent. In contrast to the halogenated deoxyuridines, catabolism of A Z T was not detected. (Methanol concentrations were increased to 20% for I U d R and 40% for AZT; detection at 280 nM was used for the halogenated derivatives.)

Cellular location of thymidine phosphorylase When whole blood was serially diluted with plasma, thymidine phosphorylase activity was directly proportional to the number of ceils present (Fig. 7a). When Triton X-100 (a non-ionic detergent) was added to blood, thymidine phosphorylase activity increased in a concentration-dependent fashion, suggesting that it is membrane-bound (Fig. 7b). Similar results were obtained with Nonidet P40 and Tween 80. A cytosolic location was also suggested by the dose-dependent inhibition of thymidine phosphorylase activity by dipyridamole, which occurred only when cells were intact (Fig. 7c). Assay after fractionation of blood cells using the Percoll mini-gradient technique, illustrated in Fig. 8, demonstrated that most thymidine phosphorylase activity co-isolated with the platelet/small lymphocyte fraction (Fig. 9). Further evidence that most blood thymidine phosphorylase activity in normal blood is located in

Fig. 8. Fractionation of blood using Percoll mini-gradients. (a) A Percoll mini-gradient loaded with blood before (right) and after (centre) centrifugation which produced 3 equal 0.7-ml fractions as illustrated by the diagram at left. Fraction 1 contains plasma but no ceils. (b) Fraction 2. The small dark cells are platelets. Small lymphocyteswhich are more refractile (arrows) from a minority population. Phase contrast. Bar = 40 /,tm. (c) Fraction 3 contained mainly red cells. A few granulocytes can also be distinguished (arrows). Phase contrast. Bar = 40/~m.

a

1 2

!

j

lem.

109

170

100

a

lysales r = 0.60

120 .>_ tl:l

O ¢.-

50

E ¢-. :,-,,

70 cells intact r = 0.77

0 1

2

3

@I

20 0.

4

fraction Fig. 9. Distribution of thymidine phosphorylase activity in blood fractions. Fractions were produced using Percoll mini-gradients as described in the text and illustrated in Fig. 8. Samples from each fraction were mixed with an equal volume of 2% w : v Triton X-100 before assay. Thymine production in each fraction is expressed as a percentage of thymine production in a lysate of unfractionated blood. The discrepancy in total activity of whole blood vs. fractions may be due to inhibitors in plasma. Fractions: 1, red cells and some leukocytes; 2, plasma, no cells; 3, platelets and some lymphocytes; 4, whole blood lysate. Serum contained about the same amount of activity as plasma.

I

0.2 0.3 thrombocrit

170

b

lysates r = 0 . 7 3

120 O C:

E r.-,

[]

0 • 0 ~

[]

•

70

~-~

-~

f

[]

cells intact r = 0.26 •

I

20 100 p l a t e l e t s was p r o v i d e d b y the close c o r r e l a t i o n b e t w e e n p l a t e l e t counts a n d t h y m i n e p r o d u c t i o n in lysates (Fig. 10a) a n d t h r o m b o c r i t a n d activity in w h o l e b l o o d b o t h b e f o r e a n d after lysis (Fig. 10b). Significant c o r r e l a t i o n s with leukocyte counts were n o t f o u n d (Fig. 10c). Quite different correlations were f o u n d in l e u k a e m i c b l o o d . T h y m i d i n e p h o s p h o r y l a s e activity was d r a m a t i c a l l y i n c r e a s e d in l e u k a e m i c c o m p a r e d to n o r m a l b l o o d samples, d e s p i t e the low p l a t e l e t c o u n t s in several of the former. T h y m i n e p r o d u c t i o n c o r r e l a t e d b e t t e r with l e u k o c y t e c o u n t s t h a n with platelet p a r a m e t e r s (Fig. l l a , b ) . These differences were statistically significant (Fig. 12).

0.4

I

I

200 300 400 [platelet](x 106ml )

500

170

C lysates r = 0.30 120

:

r-

",

j:i

E >,

~

70

c

~

l []

cells intact

20

2

.

• By n

L []

r = 0.26

I

I

4

6

•

I

8

10

[leukocyte](x 106ml ) Discussion

Results p r e s e n t e d here c o n f i r m a n d e x t e n d previous o b s e r v a t i o n s m a d e b y ourselves a n d others. T h e m a j o r difference b e t w e e n o u r p r e s e n t r e p o r t a n d earlier r e p o r t s is t h a t we have n o w b e e n a b l e

Fig. 10. Thymine production in normal blood. Correlations with haematological parameters. Whole blood samples and blood lysates from 20 normal adult donors were assayed for thymidine phosphorylase activity. Thymine production in whole blood (O) and lysates ( , ) plotted against (a) thrombocrit. (b) platelet count and (c) leukocyte count. Note that lysates contain only half the number of cells present in the corresponding whole blood samples.

110

180 •

lysales r = 0.4 []

130 ,t-"

E

e..-

8O [] []

cells inlacl r = 0.04 I ! ;

3o

20

40

60

80

I

100

£'

120

140

[platelet](x 108ml )

lao 180

lysales

-[

r = 0.56

•

e.-

"*""

80-

cells Intact r = 0.64 " ~

t=

30 0

100 6 [leukocyte](xl0 ml )

200

Fig. l l. T h y m i n e p r o d u c t i o n in l e u k a e m i c blood. C o r r e l a t i o n s with h a e m a t o l o g i c a l parameters. Symbols as in Fig. 10. T h r o m bocrits were not a v a i l a b l e for these samples.

200 lysates p < 0.005 cells intact: p < 0.05

,_._," 0~ t"-

#

100

r:.,-,.

0 INTACT

LYSED INTACT LYSED NORMAL LEUKAEMIC

Fig. 12. T h y m i d i n e p h o s p h o r y l a s e activity in n o r m a l vs. l e u k a e m i c blood samples. S h a d e d areas represent m e a n s ; uns h a d e d areas are s t a n d a r d deviations. F o r n o r m a l p o p u l a t i o n , n = 20, for leukaemics, n = 8. A single-sided S t u d e n t t test was used to c o m p a r e s a m p l e means.

to accurately standardise and quantify thymidine phosphorylase activity in whole blood and blood fractions by developing an assay which operates at saturating substrate levels. This is an important advance towards an understanding of the role of thymidine phosphorylase in a wide range of physiological and non-physiological situations. The assay we have described is rapid, simple, inexpensive, and requires no sophisticated technology apart from HPLC. It is also readily adaptable to cultured cells and to ex vivo tissues other than blood. It requires small sample volumes, and can be further scaled down (if necessary) for use in neonatal and paediatric contexts. After analysing the data obtained following assay of many normal and a few leukaemic blood samples, we have drawn conclusions which fall into 3 broad categories. First, our extensive assays of normal plasma and serum indicate that they do n o t contain concentrations of free thymidine and thymine greater than 2.5 #M, results which are in agreement with earlier reports (Hughes et al., 1973; Nottebrock and Then, 1977; Dudman et al., 1981; Holden et al., 1980; Taylor et al., 1980; Berlinger et al., 1987), although it should be noted that most of the earlier results are based on much more complex methodologies. Given that circulating thymidine is probably derived almost entirely from endogenous rather than dietary sources (Weismann, 1966; Levine et al., 1974), the (normally) low plasma concentrations found indicate that virtually all thymidine liberated from endogenous sources must be rapidly bound or metabolised. The rapid transmembranous equilibration of exogenous with intracellular thymidine (Plagemann and Wohlhueter, 1980) and the cytostatic and potentially lethal consequences of thymine nucleotide pool imbalance (Barclay et al., 1982; Meuth, 1984; Haynes, 1985; Haynes and Kunz, 1986) suggest that there is a requirement for thymidine homeostaffs in vivo. Numerous clinical studies with thymidine (comprehensively reviewed by Martin et al., 1980 and O'Dwyer et al., 1987), as well as experiments with animal models (summarized by Berlinger et al., 1987) confirm that stringent thymidine homeostasis does operate in vivo and that thymidine can be cytotoxic in vivo as well as in vitro. In the present context, thymidine toxicity

111 is especially important since peripheral blood T lymphocytes, which are routinely used in many assays designed to monitor genotoxicity, immunocompetence, allergic responses, etc. (for summary see Bodycote and Wolff, 1986), are particularly sensitive to thymidine because they contain very low levels of thymidine phosphorylase activity (Fox et al., 1979; Howell et al., 1980). Although blood thymidine phosphorylase activity is reduced at low temperatures (Fig. 6, a variable but significant amount of thymidine catabolism occurred during clotting at room temperature. The amount of thymine produced during clotting was similar to the amount produced when cell-free serum or plasma was incubated with thymidine at 37°C (Fig. 6), suggesting that temperature is an important determinant of thymidine uptake and that thymidine catabolism during clotting is due to extracellular rather than intracellular activity. In adult mammals, cell turnover is probably the main source of plasma thymidine. For example, we have estimated (see Appendix 1) that red cell production in the bone marrow could alone liberate sufficient thymidine to produce a plasma concentration at least 200-fold higher than the concentrations usually detected. Theoretically, cell turnover in peripheral tissues could also release thymidine into the blood, so the low plasma concentrations must be maintained by rapid binding, uptake or catabolism. (If the extracellular thymidine phosphorylase which we (see Fig. 6) and others - e.g., Pauly et al., 1977, 1978-have detected is not an artifact, binding may be necessary to evade catabolism.) Significant catabolism of 2'-deoxyuridine and 3 of its 5-halogenated derivatives, but not of the promising anti-HIV drug AZT, was also observed during clotting. The catabolism of the halogenated deoxyuridines by thymidine phosphorylase in blood could be predicted from previous work by Desgranges and colleagues, who have employed extensive in vitro tests using both intact platelets and the partially purified platelet enzyme to predict the pharmacological potential of a variety of substituted deoxypyrimidines (see, e.g., Desgranges et al., 1981, 1983). Their observations, together with our own, suggest that unless plasma sera are collected and stored under strictly standardised conditions, clinically important measure-

ments of serum concentrations of many antineoplastic and antiviral deoxyuridine analogues (besides those of thymidine itself) may not reflect true in vivo levels. Although previous workers (Holden et al., 1980; Dudman et al., 1981) have reported that serum thymidine phosphorylase activity is insufficient to affect thymidine concentrations during storage, they did not consider the (probably seemingly unlikely) possibility that the blood thymidine concentration may decrease during, rather than after, serum formation. The catabolism of thymidine during clotting does, however, emphasise the significance of high concentrations of thymidine in pathological sera and plasma reported here and elsewhere (Hughes et al., 1973; Holden et al., 1980; Dudman et al., 1981). Using the thymidine phosphorylase assay described here, normal basal thymidine and thymine concentrations are barely detectable, which does not exclude the possibility that they may still be considerably higher than those usually observed. The detection of measurable basal concentrations of thymidine or thymine by this assay should therefore arouse suspicion as to the existence of possible covert pathological conditions. Increased basal levels of thymidine found in leukaemic blood samples provide support for this idea. The second major point is that the characteristics of the enzyme responsible for initiating thymidine catabolism in normal human blood under the conditions of assay are those of thymidine phosphorylase rather than uridine phosphorylase, because significant catabohsm of uridine was not detected in our assays. Further work (to be published elsewhere) has shown that catabolism of thymidine in blood is inhibited in this assay by 6~aminothymine, a potent and specific inhibitor of thymidine phosphorylase (Desgranges et al., 1983). The relatively low level of thymidine phosphorylase activity in plasma and in the red blood cell/granulocyte fractions (Fig. 9) is consistent with previous reports (Pero et al., 1984; Bodycote and Wolff, 1986). Although mammalian liver, kidneys and spleen all contain substantial thymidine phosphorylase activity (Friedkin and Roberts, 1954; Friedkin and Kalckar, 1961), plasma activity is probably more likely to originate from damaged platelets (which are notoriously sensitive

112 to all types of trauma, including centrifugation), than from secretion by hepatic or other sources. Observations of high levels of thymidine phosphorylase activity in plasma obtained from tumour-bearing rodents (Pauly et al., 1978) and cancer patients (Pauly et al., 1977), in which both platelet production and turnover are often pathologically increased (Grignani et al., 1986), support this notion. Further support comes from the dramatically higher levels of thymidine phosphorylase activity and the apparently decreased ratio of extracellular:intracellular activity observed in all the (admittedly limited number of) leukaemic blood samples which we have tested. These results are consistent with the increased thymidine catabolism as indicated by increased excretion of fl-aminoisobutyric acid (the endproduct of thymidine catabolism), which has been observed in many cancer patients (for review, see Van Gennip et al., 1987). It is interesting that HIV patients and individuals at risk for HIV infection, who invariably present with haematological abnormalities (Zon et al., 1987), also show signs of increased thymidine catabolism (Borek et al., 1986). It is not presently known whether abnormalities of haemostasis and thrombosis, which are the most frequent complications of malignancy, are a cause or a consequence of the disease process (Grignani et al., 1986; Goldberg, 1987; Muszbeck, 1987). It also remains to be seen whether the human genome harbours more than one thymidine phosphorylase gene copy and how thymidine phosphorylase gene expression and enzyme activity are regulated. Genes for thymidine phosphorylase are inducible by thymidine in prokaryotes (Razzell and Casshyap, 1964); the mammalian gene(s) may likewise be inducible. If this is the case, increased levels of thymidine phosphorylase activity in leukaemia could be a direct result of increased cell turnover which clearly liberates thymidine (Holden et al., 1980; Van Gennip et al., 1987). An assay such as the one described here may therefore provide useful prognostic and/or diagnostic information in any situation in which cell turnover is increased, HIV patients and those at risk for HIV forming a particularly topical and relevant category. Thirdly, our results demonstrate conclusively

that the majority of thymidine phosphorylase activity in normal human blood is located in platelet cytosol. Early reports which detected thymidine phosphorylase activity in leukocytes (mainly granulocytes) (Marsh and Perry, 1964; Gallo and Perry, 1969; Woodman, 1979) and plasma (Pauly, 1977, 1978) appear to have overlooked platelet contamination or lysis as a possible source of activity. The authors of these reports were unable, or did not attempt, to express thymidine phosphorylase activity in terms of the total amount present in whole blood. More recent reports (Pero et al., 1984; Shaw, 1986; Bodycote and Wolff, 1986) concur with our finding that a major proportion of total blood thymidine phosphorylase activity is located in platelets. Apart from platelets, small lymphocytes were the major cell type present in the blood fraction which contained most thymidine phosphorylase activity. T lymphocytes, which are known not to contain significant thymidine phosphorylase activity (Fox et al., 1979; Howell et al., 1980), constitute a large fraction of this cell population (Hoffbrand and Pettit, 1984). In addition, we have estimated (see Appendix 2) that even the leukocytes which do contain thymidine phosphorylase are unlikely to contribute much to the total blood activity under normal circumstances, although this situation clearly does not pertain to leukaemic blood. Inhibition by dipyridamole of thymidine phosphorylase activity in intact cells, but not in lysates (Fig. 7c), and also the increase in activity which accompanies detergent-mediated permeabilisation and cell lysis (Fig. 7b) confirms work by Desgranges et al. (1981) which suggested that platelet thymidine phosphorylase has a cytosolic rather than membrane or organellar location. The (normally) intraplatelet location for most thymidine phosphorylase is substantiated by the close correlation between enzyme activity and platelet parameters. The close correlation between platelet counts and thymidine phosphorylase activity in blood lysates than between platelet counts and whole-blood activity with cells intact (Fig. 10a) provides further evidence that the enzyme is intracellular. The extremely high platelet surface area:volume ratio (see Appendix 2) would ensure that thymidine phosphorylase could be made rapidly available when required, as well as providing opportunities

113

for far subtler forms of regulation than would otherwise be possible. Our observation that serum and plasma levels of thymidine phosphorylase are not significantly different suggests that platelet thymidine phosphorylase is not secreted in response to platelet agonists, although secretion followed by inactivation in the extracellular milieu remains a possibility which we are currently investigating. Human blood (platelet) thymidine phosphorylase has the capacity to handle an extraordinarily large nucleoside load relative to apparent normal physiological requirements. However, it is not difficult to envisage that very high thymidine concentrations might be generated locally at sites of rapid cell turnover, regardless of whether such turnover is due to physiological or pathological processes (e.g., red cell production and leukaemia respectively). We therefore suggest: (i) the seriousness of the genetic consequences of both thymidine starvation and overload indicates a need for peripheral as well as central (i.e., hepatic and renal) control of available thymidine; (ii) thymine and thymidine concentrations in normal human plasma and serum are in the micromolar range, although in vivo concentrations may be higher than usually reported, and very much higher concentrations are likely to occur at localised sites in vivo; (iii) thymidine phosphorylase activity in human peripheral blood is located mainly in platelet cytosol and has a potential capacity for thymidine catabolism which appear to be far in excess of the normal physiological requirement; (iv) several 5substituted deoxyuridine analogues are susceptible to catabolism by blood (platelet) thymidine phosphorylase, but the anti-HIV drug AZT is not, which is consistent with what is known about the pharmacokinetics of these drugs (Myers, 1981; Klecker et al., 1986); (v) blood thymidine phosphorylase activity in vitro (and by extension, in vivo) may be regulated at the membrane level as well as intracellularly; (vi) provided that an adequate concentration of deoxyribonucleoside donor(s) is available, thymidine phosphorolysis is reversible in vitro, and presumably also in vivo; (vii) the importance of the suggested physiological role of platelet thymidine phosphorylase (i.e., regulation of plasma thymidine) cannot be too highly

stressed, since by controlling the availability of thymidine it may exert a ubiquitous and potentially crucial influence on whole-body nucleotide metabolism; finally, (viii) the well-documented aggregatory and adhesive potential which platelets develop in response to numerous exogenous stimuli suggests that platelet thymidine phosphorylase activity could be targeted to specific 'crisis' sites (e.g., wounds or tumours), where nucleotide pool imbalances could be expected and where local conditions could favour either thymidine catabolism or thymidine regeneration from thymine and deoxynucleoside donors. At these sites alterations in cellular thymine nucleotide pools could either stimulate or arrest cell proliferation, depending on pre-existing nucleotide pool balance. This platelet function is probably at least equal in biological significance to the much-studied and publicised (but largely speculative) roles in carcinogenesis and wound repair which have been proposed for various peptide growth factors of platelet origin (see, e.g., Ross et al., 1986; Deuel, 1987). Platelets avidly take up and phosphorylate purine bases and nucleosides; indeed, the involvement of platelets in urine nucleotide metabolism has been extensively studied for more than 20 years (for reviews, see Holmsen, 1985 and Ashby, 1987). Our data indicate that regulation of blood thymidine is also an important, although poorly appreciated, platelet function. Taken together, the data suggest an important, perhaps vital, role for platelets in maintaining nucleotide balance; they further imply that megakaryocytes (the precursors of platelets) (see Hoffbrand and Pettit, 1984) might perform an analogous function within the bone marrow. The intracellular location and wide variation between individuals (also reported by Bodycote and Wolff (1986)) in blood thymidine phosphorylase activity suggests that control of its activity is complex and worthy of investigation. Since this work was completed, we have used the assay described here to investigate the effects on blood thymidine phosphorylase activity of several platelet agonists and antagonists besides dipyridamole. The results, which we believe can provide important insights into both platelet function and regulation of cell proliferation in general, will be reported separately.

114 Appendix 1. Calculation of thymidine liberated as a result of red-cell turnover a

Human (diploid) cell nuclei contain about 8 pg of DNA, which has a (G + C) content of roughly 40%, thus the thymidine content is about 1.2 pg per nucleus. The normal human adult red blood cell count is about 5 x 1012 1-1 and the average red-cell lifespan is 120 days, so the production rate must be about (5 x 1012/120) cells per day per litre of blood. In adults, the average circulating blood volume is about 5 litres, so degenerating erythroblast nuclei would yield about 250 mg (5 x 5 X 1012X 1.1/120 pg) of thymidine per day, which would have the potential to elevate the peripheral blood thymidine concentration at the rate of about 0.2 mM per day (8 /~M h - l ) . If confined to the bone marrow (the total volume of which is less than 200 ml), the daily concentration increment would be in the region of 5 mM. Even greater concentrations would occur in the absence of rapid transmembrane equilibrium. Although some thymidine must be recycled during erythropoiesis, it is known that DNA synthesis in the maturing erythrocyte series depends increasingly on de novo synthesis of thymidine rather than salvage (DiSrmer et al., 1972). Other cells must therefore play some part in thymidine homeostasis. Appendix 2. Relative platelet and leukocyte contributions to whole-blood thymidine phosphorylase activity a

Our experimental data show that peripheral blood thymidine phosphorylase normally operates at substrate concentrations which are orders of magnitude below saturation and at which membrane transport is likely to a significant determinant of activity. Some idea of the relative contributions to total thymidine phosphorylase activity under physiological conditions might therefore be gained from a comparison of cell surface areas. To simplify calculations, platelets and leukocytes are assumed to be spherical. (The surface area of a sphere of radius r is 47rr2).

a Haematologicaldata from Hoffbrand and Pettit (1984).

The mean platelet volume in normal adults is about 9/~m 3 (range, 7.4-10.4/~m3). The surface of a sphere having this volume is about 21/Lm 2. The average blood platelet concentration in normal adults is 2.75 x 10 8 (range, 1.5-4.0 x 10 8) m1-1 which would provide a total surface area of about 5.8 x 10 9/~m 2 m1-1. Circulating leukocytes have an average radius of about 5 #m, giving a surface area of about 314 Fm 2. The leukocyte concentration in normal adults is 7.5 x 10 6 (range 4.0-11.0 x 10 6) m1-1, providing a surface area of 2.4 x 10 9 p m 2 ml -a, however, only about half this population (the granulocytes) contain significant thymidine phosphorylase activity (see references in text). Even allowing for this, the attribution of roughly 3 / 4 of total thymidine phosphorylase activity to platelets is still likely to be an underestimate, because platelets are not spherical and their effective surface area is increased several times by the presence of complex internal cannalicular system which is surface-connected. Acknowledgements

We thank students and colleagues who donated blood, Dr C. Rana and Sr Louise Crellin of La Trobe University Health Service and Sr Sue Bevan form the Dept. of Genetics and Human Variation for assistance with blood collection and provision of haematological data. Thanks are also due to Dr Annabelle Duncan and Mrs Annetta Miller for their invaluable secretarial, technical and organisational skills, to Mr Russell Thompson for obtaining reference material and Miss Wendy Leong for proofreading. References

Ashby, B. (1987) Nucleotide metabolism, in: H. Holmsen (Ed.), Platelet Responses and Metabolism, Vol. II, CRC, Boca Raton, FL, pp. 215-231. Barclay, B.J., B.A. Kunz, J.G. Little and R.H. Haynes (1982) Genetic and biochemical consequences of thymidylate stress, Can. J. Biochem.,60, 172-194. Barlow, S., and M.G. Ord (1975) Thymidine transport in phytohemagglutinin-stimulated pig lymphocytes, Biochem. J., 148, 295-302. Berlinger, W.G., R.A. Stene, R. Spectorand A.S. al-Jurf (1987) Plasma and cerebrospinal fluid nucleoside and oxypurines in acute liver failure, J. Lab. Clin. Med., 110, 137-144.

115 BjurseU, B., and L. Skoog (1980) Control of nucleotide pools in mammalian cells, Antibiotics Chemother., 28, 78-85. Bodycote, J., and S. Wolff (1986) Metabolic breakdown of [3H]tbymidine and the inability of measure human lymphocyte prohferation by incorporation of radioactivity, Proc. Natl. Acad. Sci. (U.S.A.), 81, 4749-4753. Borek, E., O.K. Sharma, F.L. Buschman, D.L. Cohn, K.A. Penley, F.N. Judson, B.S. Bobozin, C.R. Hornsburgh and C.H. Kirkpatrick (1986) Altered excretion of modified nucleotides and beta-aminoisobutyric acid in subjects with Acquired Immunodefficiency Syndrome or at risk for Acquired Immunodeficiency Syndrome, Cancer Res., 46, 2551-2561. Cooper, G.M., W.F. Dunning and S.F. Greer (1972) Role of catabolism in pyrimidine utilization for nucleic acid synthesis in vivo, Cancer Res., 32, 390-397. Cool~er, R.A., S. Perry and T.R. Breitman (1966) Pyrimidine metabolism in human leukocytes, I. Contribution of exogenous thymidine to DNA-thymine and its effect on thymine, Cancer Res., 26, 2267-2275. Dawson, R.M.C., D.C. Elliott, W.H. Elliott, and K.M. Jones (1986) Data for biochemical Research, 3rd edn., Clarendon, Oxford, pp. 103-114. Desgranges, C., G. Razaka, M. Rabaud and H. Bricaud (1981) Catabolism of thymidine in human blood platelets, purification and properties of thymidine phosphorylase, Biochim. Biophys. Acta, 654, 211-218. Desgranges, C., G. Razaka, M. Rabaud, H. Bricaud, J. Balzarini and E. de Clercq (1983) Phosphorolysis of (E)-5-(2bromovinyl)-2'-deoxyuridine (BVDU) and other 5-substituted-2'-deoxyuridines by purified human thymidine phosphorylase and intact blood platelets, Biochem. Pharmacol., 32, 3583-3590. Desgranges, C., E. de Ciercq, G. Razaka, F. Drouillet, I. Belloc and H. Bricaud (1986) Deoxyribosyl exchange reactions leading to the in vivo regeneration of the antiviral agents (E)-5-(2-bromovinyl)-2 '-deoxyuridine, 5-ethyl-2'-deoxyuridine and 5-(2-chloroethyl)-2'-deoxyuridine, Biochem. Pharmacol., 35, 1647-1653. Deuel, T.F. (1987) Polypeptide growth factors: roles in normal and abnormal-cell growth, Annu. Rev. Cell Biol., 3, 443. Drrmer, V.P., W. Brinkmann and P. Dahr (1972) Thymidylatde novo-Synthese, salvage pathway und DNS-Synthese wiihrend der Differenzierung von Erythroblasten, Blut, 25, 185-189. Dudman, N.P., W.B. Deveski and M.H.N. Tattersall (1981) Radioimmunoassay of plasma thymidine, uridine, deoxyuridine and cytidine/deoxycytidine, Analyt. Biochem., 115, 428-437. Ensminger, W.D., and E. Frei (1977) The prevention of methotrexate toxicity by thymidine infusions in humans, Cancer Res., 37, 1857-1863. Ensminger, W.D., and E. Frei (1978) High-dose intravenous and hepatic artery infusions of thymidine, Clin. Pharmacol. Ther., 24, 610-615. Fox, R.M., S.K. Piddington and A.A. Piper (1979) Thymidine phosphorylase deficiency in cultured human T leukemic lymphocytes: determinants of sensitivity of pyrimidine antimetabolites, Proc. Am. Assoc. Cancer Res., 20, 262.

Friedkin, M., and DeW. Roberts (1954) The enzymatic synthesis of nucleosides, I. Thymidine phosphorylase in mammalian tissue, J. Biol. Chem., 207, 245-256. Friedkin, M., and H. Kalckar (1961) Nucleoside phosphorylases, in: P.D. Boyer, H. Lardy and D. Myrback (Eds.), The Enzymes, Vol. 5, 2nd edn., Academic, New York, pp. 237-255. Gallo, R.C., and S. Perry (1969) The enzymatic mechanisms for deoxythymidine synthesis in human leukocytes, IV. Comparisons between normal and leukemic leukocytes, J. Clin. Invest., 48, 105-116. Goldberg, R.J. (1987) Venous thromboembolism and malignancy (Editorial), Arch. Intern. Med., 147, 1893-1894. Grignani, G., L. Pacchiarini and M. Pagliarino (1986) The possible of blood platelets in tumor growth and dissemination, Haematologica, 71,245-255. Hartwick, R.A., S.P. Assenza and P.R. Brown (1979a) Identification and quantitation of nucleosides, bases and other UV-absorbing compounds in serum, using reversed-phase high-performance liquid chromatography, I. Chromatographic methodology, J. Chromatogr., 186, 647-658. Hartwick, R.A., A.M. Krstulovic and P.R. Brown (1979b) Identification and quantitation of nucleosides, bases and other UV-absorbing compounds in serum, using reversedphase high-performance chromatography, II. Evaluation of human sera, J. Chromatogr., 186, 659-676. Haynes, R.H. (1985) Molecular mechanisms in genetic stability and change: the role of deoxyribonucleotide pool balance, in: F.J. de Serres (Ed.), Genetic Consequences of Nucleotide Pool Imbalance, Plenum, New York, pp. 1-123. Haynes, R.H., and B.A. Kunz (1986) The influence of thymine nucleotide depletion on genetic stability and change in eukaryotic cells, Curr. Sci., 55, 1-11. Hoffbrand, A.V., and J.E. Pettit (1984) Essential Haematology, 2nd edn., Blackwell, Oxford. Holden, L., A.V. Hofbrand and M.H.N. Tattersall (1980) Thymidine concentrations in human sera: variations in patients with leukaemia and megaloblastic anaemia, Eur. J. Cancer, 16, 115-121. Holmsen, H. (1985) Nucleotide metabolism of platelets, Annu. Rev. Physiol., 47, 677-690. Howell, S.B., R. Taetle and J. Mendelsohn (1980) Thymidine as a chemotherapeutic agent: sensitivity of normal human marrow, peripheral blood T cells, and acute nonlymphocytic leukemia, Blood, 55, 505-510. Hughes, W.L., M. Christine and B.D. Stollar (1973) A radioimmunoassay for measurement of serum thymidine, Anal. Biochem., 55, 468-478. Klecker, R.W., J.M. Collins, R. Yarchoan, R. Thomas, J.F. Jenkins, S. Broder and C.F. Myers (1986) Plasma and cerebrospinal fluid pharmacokinetics of 3'-azido-3'-deoxythymidine: a novel pyrimidine analog with potential application for the treatment of patients with AIDS and related diseases, Clin. Pharmacol. Ther., 41,407-412. Kornberg, A. (1980) DNA Replication, W.H. Freeman, San Francisco, CA, pp. 39-85. Kraut, A., and E.W. Yamada (1971) Cytoplasmic uridine phosphorylase of rat liver, J. Biol. Chem., 246, 2021-2030. Kufe, D.W., P. Beardsley, D. Karp, L. Parker, A. Rosowsky,

116 G. Caneilos and E. Frei (1980) High-dose thymidine infusions in patients with leukemia and iymphoma, Blood, 55, 580-589. Levine, R.L., N.J. Hoogenraad and N. Kretchrner (1974) A review: Biological and clinical aspects of pyrimidine metabolism, Pediatr. Res., 8, 724-734. Marsh, J.C., and S. Perry (1964) Thymidine catabolism by normal and leukemic human leukocytes, J. Clin. Invest., 43, 267-278. Martin, D.S., R.L. Stolfi, R.CV. Sawyer, R. Nayak, S. Spiegelman, C.W. Young and T. Woodcock (1980) An overview of thymidine, Cancer, 45, 1117-1128. Meuth, M. (1984) The genetic consequences of nucleotide precursor pool imbalance in mammalian cells, Mutation Res., 126, 107-112. Muszbek, L. (Ed.) (1987) Hemostasis and Cancer, CRC, Boca Raton, FL. Myers, C.E. (1981) The pharmacology of the fluorinated pyrimidines, Pharmacol. Rev., 33, 1-15. Nottebrock, H., and R. Then (1977) Thymidine concentrations in serum and urine of different animal species and man, Biochem. Pharmacoi., 31, 1857-1861. O'Dwyer, P.J., S.A. King, D.F. Hoth and B. Leyland-Jones (1987) Role of thymidine in biochemical modulation: a review, Cancer Res., 47, 3911-3919. Pauly, J.L., M.G. Schuller, A.A. Zelcer, T.A. Kirss, S.S. Gore and M.J. Germain (1977) Quantification and comparative analysis of thymidine phosphorylase in plasma of healthy subjects and cancer patients, J. Natl. Cancer Inst., 58, 1587-1590. Pauly, J.L., N.S. Paolini, R.L. Ebarb and M.J. Germain (1978) Elevated thymidine phosphorylase activity in the plasma and ascitic fluids of tumor-beating animals, Proc. Soc. Exp. Biol. Med., 157, 262-267. Pero, R.W., D. Johnson and A. Olsen (1984) Catabolism, of exogenously supplied thymidine to thymine and dihydrothymine by platelets in human peripheral blood, Cancer Res., 44, 4955-4961. Plagemann, P.G.W., and R.M. Wohlhueter (1980) Permeation of nucleosides, nucleic acid bases and nucleotides in animal cells, Curr. Top. Membr. Transp., 14, 225-230. Puzas, J.E., and J.S. Brand (9186) The effect of bone cell stimulatory factors can be measured with thymidine incorporation only under specific conditions, Calcif. Tissue Int., 39, 104-108. RazzeU, W.E., and P. Casshyap (1964) Substrate specificity and induction of thymidine phosphorylase in Escherichia coil, J. Biol. Chem., 239, 1789-1793.

Reichard, P. (1978) From deoxynucleotides to DNA synthesis, Fed. Proc., 37, 9-14. Ross, R., E.W. Raines and D.F. Bowen-Pope (1986) The biology of platelet-derived growth factor, Cell, 46, 155-169. Rubini, J.R., E.P. Cronkite, V.P. Bond and T.M. Fleidner (1960) The metabolism and fate of tritiated thymidine in man, J. Clin. Invest., 39, 909-918. Schneider, W., C. Doenecke and P.G. Scheurlen (1973) Detection of thymidine-phosphorylase in human blood platelets, Klin. Wschr., 51, 287-288. Shaw, T. (1986) Rapid and complete degradation of thymidine by peripheral blood platelets: implications for genotoxicity assays. Thrombosis Res., Suppl. VI, 16 (abstr. 326). Shaw, T., and D.G. MacPhee (1986) Rapid and complete degradation of thymidine by peripheral blood platelets: implications for genotoxicity assays, Mutation Res., 163, 75-85. Simpson, R.C., and P.R. Brown (1986) High-performance liquid chromatographic profiling of nucleic acid components in physiological samples, J. Chromatogr., 379, 269-311. Steinberg, J.A., M. Otten and G.B. Grindley (1979) Isolation of a DNA polymerase-associated regulatory protein from calf thymus, Cancer Res., 39, 4330-4335. Taylor, G.A,, P.J. Dady and K.R. Harrap (1980) Quantitative high-performance liquid chromatography of nucleosides and bases in human plasma, J. Chromatogr., 183, 421-431. Usher, D.C., and H. Reiter (1977) Catabolism of thymidine during the lymphocyte cell cycle, Cell, 12, 365-370. Van Gennip, A.H., E.J. Van Bree-BIom, N.G.G.M. Abeling, A.J. Van Erven and P.A. Voute (1987) Beta-aminoisobutyric acid as a marker of thymine catabolism in malignancy Clin. Chim. Acta, 165, 365-378. Vincent, R., and D. Nadeau (1984) Adjustment of the osmolarity of Percoll for the isopycnic separation of cells and cell organelles, Anal. Biocbem., 141, 322-328. Weissmann, S.M. (1966) Human pyrimidine metabolism, J. Am. Med. Assoc., 195, 117-120. Woodman, P. (1979) Thymidine phosphorylase activity in plasma: a cancer marker or an artifact of filtration?, Proc. Soc. Exp. Biol. Med., 162, 175-178. Zimmerman, M. (1964) Deoxyribosyl transfer, II. nucleoside: pyrimidine deoxyribosyl transferase activity of three partially purified thymidine phosphorylases. J. Biol. Chem., 239, 2622-2627. Zon, L.I., C. Arkin and J.E. Groopman (1987) Haematologic manifestations of the human immune deficiency virus (HIV), Br. J. Haematol., 34, 251-256.